Road Profiler User Group (RPUG) Atlanta, Georgia

Ministry of TransportationMinistère des Transports

Development of A New Pavement Distress Evaluation Guide for Ontario

Ministry of Transportation

Road Profiler User Group (RPUG)Atlanta, Georgia

December 10, 2009

Li Ningyuan, Ph.D., P.Eng. Senior Pavement Management Engineer


Presentation Outlines

• MTO Current Practice in Pavement Evaluation• Issues on Subjective Assessment of Pavement

Distresses (14~16 Individual Distresses)• Studies on Rationalizing Evaluation of

Pavement Distresses at Network Level• Conclusions and Summary • Move Forward Semi-Automation/Automation of

Pavement Distress Assessment


Current MTO Practice in Pavement Data Collection

• IRI (automated process since 1997)• Rutting (automated process since 1997)• DMI (field visual assessment since 1980)Subjective evaluation of pavement condition

using Distress Manifestation Index (DMI)

• PCI (Pavement Condition Index) is a function of IRI and DMI


DMI Background

• Distress Manifestation Index (DMI) is used to indicate an evaluator’s overall assessment of a pavement condition based on his/her survey in the field

• DMI is a subjective evaluation of pavement condition rated by trained and experienced pavement evaluators

• DMI represents evaluation of an individual pavement section, typically 10 kilometers in section length and uniform performance

• MTO’s Pavement Condition Rating Manuals


Pavement Condition Rating Manuals


Sample Distress Attributes


Typical Pavement Distresses


Overview of Old System


DMI Components

• There are between 13 and 16 individual distress components depending on pavement type, which is grouped into three categories:−Deformation−Surface Distresses−Cracks


i = distress type iwi = weighting factor assigned to distress isi = severity of distress i ei = extent of distress i

The scale of DMI is ranged from 0 to 10 in MTO PMS

Distress Manifestation Index (DMI)

DMI = Σwi (si+ei)15

i=1

Asphalt Concrete Pavement(AC)

Weight(Wi)

Ravelling and Coarse Aggregate Loss 3Flushing 1.5Rippling and Shoving 1Wheel Track Rutting 3Distortion 3Longitudinal Wheel Track: Sing. / Multi. 1.5Longitudinal Wheel Track: Alligator 3Longitudinal Meandering and Midlane 1Transverse: Half, Full and Multiple 1Transverse: Alligator 3Centreline: Single and Multiple 0.5Centreline: Alligator 2Pavement Edge: Single and Multiple 0.5Pavement Edge: Alligator 1.5Random/Map 0.5

Portland Cement Concrete(PCC)

Weight(Wi)

Ravelling and Coarse Aggregate Loss 0.5Polishing 1.5Scaling 1.5Potholing 1Joint and Crack Spalling 2Faulting 2.5Distortion 1Joint Failure 3Longitudinal Joint Separation 1Longitudinal and Meandering Cracking 2Transverse Joint Creep 0.5Transverse Cracking 2Joint Sealant Loss 0.5Diagonal Corner and Edge Crescent 2.5“D” Cracking 3

Weights of Individual Distresses

Weights of Individual Distresses (Cont’)Composite Pavement

COM Weight

(Wi)Ravelling and Coarse Aggregate Loss 3Flushing 1.5Spalling 2Tenting/Cupping 2.5Wheel Track Rutting 3Joint Failures 3Distortion and Settlement 1Longitudinal Meandering (Single & Multiple) 2Transverse: Single 1Transverse: Multiple 1Transverse Joints: Sawed 0.5Transverse Joints: Reflective 2Centreline: Single 0.5Centreline: Multiple 1.5Diagonal, Corner and Edge Crescent 2.5Random/Map 0.5

Surface TreatedST

Weight(Wi)

Cover Aggregate Loss 3

Flushing 2

Streaking 1

Potholing 1

Rippling and Shoving 2

Wheel Track Rutting 3

Distortion 3

Longitudinal Cracking 1

Transverse Cracking 0.5

Pavement Edge Break 2

Pavement Edge Cracking 1

Alligator Cracking 3


Scale for Severity Rating

• Five (5) severity Levels for AC, PCC and COM type of pavement, ranging from 1 to 5

• Three (3) severity Levels for ST pavement

( )

=====

=

54321

,

,

nnnnn

VerySevereSevereModerateSlightVerySlight

nSi

( )

===

=

432

,nnn

SevereModeratelight

nSi


Scale for Density Rating

• Five (5) Density/Extent levels for AC, PCC and COM pavement, in terms of percentage %

• Three (3) Density/Extent levels for ST pavement

( )

=====

−−−−−

=

54321

%,10080%,8060%,6040%,4020

%,200

nnnnn

nDi

( )

===

−−−

=432

%,10050%5020

%,200

nnn

nDi


Issues Concerned with DMI

• Safety and Accessibility in Field OperationImpossible to operate on high-volume traffic freewaysRestricted by vehicle operation speed and visual assessment

• Quality and Productivity of Data CollectionPoor quality (accuracy, repeatability and accountability)Subjective bias and errors in evaluation

• Impacts on Pavement Management ProcessPerformance evaluation and needs analysisDecisions in selecting pavement treatmentsMaintenance programming and investment planning


Facts Considered in Reducing Some individual Distresses• Type and number of distresses• Weighting factors of individual distresses• Distress severity and density• Minimize impacts on current practice• Preserve historic data and consistency• Six years of data were used in performing this

study, covering four pavement types, AC, PCC, COM, ST


Test Design and Analysis

• Group distress and re-define DMI(#) • DMI(#) stands for DMI that is calculated by use of the

existing formula but excluding individual distresses that have weight factors lower than # −DMI(1) contains distresses with weight >=1−Similar definition for DMI(1.5), DMI(2) and DMI(3)

• DMI(C) and DMI(C&R) include only cracking / cracking & rutting as distresses

• DMI (T) is the original DMI (including all distresses)


• Use six years historic data (from 2001 to 2006) extracted form MTO pavement management databases

• DMI (#) were calculated and then compared with the DMI (T) to produce an error percentage

• Note that the sample size varies significantly between the four pavement types: ―1344 AC Sections, 26 PCC Sections, 22 COM

Sections, and 271 ST Sections

Test Design and Analysis (Cont’s)

Analysis Results for AC Pavements

AC % Errors

-10%

-5%

0%

5%

Avg DMI(1) Avg DMI(1.5) Avg DMI(2) Avg DMI(3) Avg DMI(C ) Avg DMI(C&R)

AC 01AC 02AC 03AC 04AC 05AC 06

Comparison between errors from each modified DMI (AC)


ST % Errors

-10%

-5%

0%

5%

10%

15%

20%

25%

Avg DMI(1) Avg DMI(1.5) Avg DMI(2) Avg DMI(3) Avg DMI(C ) Avg DMI(C&R)

ST 01ST 02ST 03ST 04ST 05ST 06

Comparison between errors from each modified DMI (ST)

Analysis Results for ST Pavements

Remaining Distress Components of DMI(2) for AC Pavement

Asphalt Concrete Pavement (AC) Weight (Wi)Ravelling and Coarse Aggregate Loss 3Flushing 1.5Rippling and Shoving 1Wheel Track Rutting 3Distortion 3Longitudinal Wheel Track: Sing. / Multi. 1.5Longitudinal Wheel Track: Alligator 3Longitudinal Meandering and Midlane 1Transverse: Half, Full and Multiple 1Transverse: Alligator 3Centreline: Single and Multiple 0.5Centreline: Alligator 2Pavement Edge: Single and Multiple 0.5Pavement Edge: Alligator 1.5Random/Map 0.5

Asphalt Concrete Pavement (AC) Weight (Wi)

Ravelling and Coarse Aggregate Loss 3Wheel Track Rutting 3Distortion 3Longitudinal Wheel Track: Alligator 3Transverse: Alligator 3Centreline: Alligator 2


Frequency of Distresses

Long. S & M

Long. Alg

Long. M & M Transverse Alg.

Centreline S & M

Centreline Alg.

PavEdge S&M

PavEdge Alg.

Random/Map

Transverse HFM

Ravelling & C.A Loss

Flushing

Rippling and Shoving

Distortion

Wheel Track Rutting


List of Remaining Distresses# Distress Name1 Potholing2 Ravelling and C. A. Loss3 Longitudinal Cracking4 Transverse Cracking5 Map/Alligator Cracking

• Taking weighting and frequency into account, a list of proposed distresses was made

• Using these distresses calculations show that a DMI calculated with only these values is accurate to 1.5% in flexible pavement


Network Level Program

• Collection of:• Transverse Profile (Rutting)• Longitudinal Profile (Roughness)• Orientation (Crossfall) • GPS data

• Over 1390 PMS Sections• North and West on non-divided highways• Both directions on freeways


ARAN Data Delivery

• Currently in Excel Format• Main delivery at year end

• CD to each region with all pavement data incl:• 50m Detailed Data• Chart File• Summary sheet

• *New for 2007 plotted network/monitor data by region/district

* Still have to determine best method of delivery


Sony HDTV Video Camera


Video Collection• Images captured by a Sony HDTV wide angle camera.

• Great image quality and contrast• 1920 x 1080 resolution• 16:9 aspect ratio• 90 degree field of view

• Camera mounted inside ARAN cabin• This allows environmental control

• Capture Interval is 5 meters, total storage per year ~ 2 TB• Network Level: Primary Highways, North/West bound, Lane 2• Bidirectional information acquired from all of Central, Southern &

Eastern Ontario– Also where possible in Northern and Northwestern Ontario

• Project Level testing - as requested


Quality Examples - Acceptable


Proposed Method

• ARAN will continue to collect HD video-logs

• Raters will then be able to play these video-logs at the office, and manually evaluate roads from images

• This would take care of safety and accessibility issues for high-volume roads, as well as allow raters to make more accurate evaluations


Suggested Implementation• Images must be viewed on large monitors to

ensure that the resolution of HD images is not lost

• Video software must be installed, and videos can be viewed on a secondary monitor so that section data and other information can be viewed on the first


Potholing


Coarse Aggregate Loss


Ravelling


Ravelling 2


Longitudinal Cracking


Transverse Cracking


Transverse Cracking 2


Alligator Cracking


More Benefits of Video-logging

• Repeatability is also increased, as any video-log can be revisited for verification purposes

• Video logs can be paused, rewound, and even played backwards at any time for further inspection

• Images are tied to GPS, roughness and retting, and positioning data allowing raters to return to areas of high distress as necessary


Next Steps

• Determine severity and density levels from video-logs

• Literature review of current evaluation methods• Develop evaluation procedure methods• Timing, follow-up

• [FLOW CHART]


Next Steps

Introduce Video-Logs

Define levels

Gather images of chosen distresses

Sample roads and compare ratings

Reduce list of distresses

Compile levels into procedure

Train regions and gather feedback

Implement Video-Log Evaluations

Current Stage

Acceptable

Needs revision

Inaccurate

Accurate


Identifying Severity & Density Levels

• Images of different severity and density levels will be gathered from video-logs

• Corresponding level guidelines will then be made for raters

• The number of severity levels will be decreased to 3 to make evaluations easier


Study of Current Evaluation Methods

• A more thorough study of current evaluation methods will be done in order to determine how video-logging can be incorporated in the future

• There is no standard definitions for severity, so an inter-region group study will be carried out for determining bench marks


Developing Evaluation Procedure

• By combining the new distress list with severity levels, current evaluation methods, and video-logging, an evaluation procedure will be developed

• Once the procedure is developed, training can be implemented and tested for quality by comparing manual ratings from roadside evaluations and manual ratings from video-logs


Following up

• As distresses are chosen, current PDDC data will be compared to data collected from video-logging evaluations

• This will be followed up by gathering feedback from regions to ensure standard methods are used within MTO


Benefits

• In-office view of highway network• reduction of field trips • faster response to inquiries

• Historical record of highway conditions• Validation/QA of field data• Roadside asset inventory data collection


Li Ningyuan, P.Eng.Tel: 416-235-3518

[email protected]

Thank You!

Documents

Road Profiler User Group (RPUG) Atlanta, Georgia